Terence Tao

Mathematician and analyst of AI's impact on science.

Claims

As speaker

Mentioned in

Key positions and views

The future of mathematics is hybrid human-AI collaboration, where AI's breadth in generating ideas complements human experts' depth in complex reasoning (Claims 5, 6).

AI has fundamentally shifted the scientific bottleneck from hypothesis generation, which is now nearly free, to the verification and evaluation of a massive number of new theories (Claims 4, 17).

Current AI architectures are not on a direct path to replacing human mathematicians due to fundamental limitations like the inability to build on progress cumulatively and a low success rate on novel problems (Claims 1, 10, 19).

AI will automate the bulk of routine mathematical work within a decade, transforming the day-to-day activities of professional mathematicians (Claim 2).

By 2026, AI will be a 'trustworthy co-author' in mathematics, provided it is used correctly by human experts who can guide and verify its output (Claim 7).

Podcast consensus on Tao

Points of consensus

▶Terence Tao consistently views AI as a powerful tool for the 'experimental' or 'idea generation' phase of science, drastically lowering the cost of producing hypotheses and enabling large-scale data gathering on technique effectiveness (Claims 3, 4, 17).Apr 2026

▶Tao believes the future of mathematics lies in hybrid human-AI teams, where human depth complements AI's breadth, asserting that a full replacement of human mathematicians is not imminent and would require new breakthroughs (Claims 5, 6).Apr 2026

▶Tao maintains that current AI models have significant limitations for core mathematical work, including a very low success rate on novel problems (1-2%), an inability to build on progress cumulatively, and a failure to speed up the most difficult parts of problem-solving (Claims 1, 10, 19).Apr 2026

Points of debate

▶There is a tension between Tao's assertion that he still uses pen and paper for the hardest problems and Nathan Labenz's report that an AI solved a challenging problem posed by Tao in weeks, a task that took humans 18 months to make progress on (Claims 1, 8).Apr 2026

▶Tao highlights both a recent surge and a subsequent pause in AI's ability to solve Erdős problems, noting that after solving about 50, progress has stalled as the 'low-hanging fruit' has been picked, suggesting a potential performance plateau (Claims 12, 21).

▶Tao predicts AI will automate the bulk of mathematicians' work within a decade, yet simultaneously states that current AI tools have only a 1-2% success rate in systematic studies, indicating a significant performance gap that must be closed to realize his prediction (Claims 2, 19).Apr 2026

▶While Tao celebrates that AI has driven the cost of idea generation to near zero, he also warns this has created a new bottleneck in verification and is flooding scientific journals with low-quality, AI-generated papers (Claims 4, 17, 20).Apr 2026

Key themes

▶The Shifting Role of the Human MathematicianApr 2026

Tao argues that AI is poised to automate the bulk of routine mathematical tasks within a decade. This will shift the human role away from laborious execution and toward verification, evaluation, and tackling the core, difficult parts of problems where AI currently fails.

The value of mathematical talent will increasingly be found in creative problem-framing and deep, intuitive understanding—skills currently beyond AI's capabilities—representing a new frontier for human capital investment and education.

▶AI as Breadth Engine, Humans as Depth EngineApr 2026

Tao characterizes AI's strength as breadth—generating thousands of ideas and exploring vast datasets for patterns. In contrast, he sees human experts as excelling at depth—cumulative, focused reasoning on a single, complex problem. He believes the most effective approach for the foreseeable future is a hybrid model that combines these complementary skills.

Organizations should focus on building collaborative workflows and tools that integrate human experts with AI systems, rather than pursuing a strategy of full human replacement, to maximize research and development outcomes.

▶The Bottleneck Shift from Generation to VerificationApr 2026

According to Tao, AI has driven the cost of generating scientific ideas and theories to near zero. This has created a new scientific bottleneck: the overwhelming task of verifying and evaluating the flood of AI-generated hypotheses, a problem exacerbated by journals being inundated with AI-generated papers.

A significant market opportunity exists for developing advanced AI-powered verification and validation tools that can sift through and rigorously test the massive output of generative models.

▶Current AI's Architectural LimitationsApr 2026

Tao points out fundamental weaknesses in current AI paradigms, such as their low success rate (1-2%) on novel math problems and their inability to build on progress cumulatively, as they 'forget' previous interactions in a new session. He believes a full replacement of human mathematicians will require breakthroughs beyond these existing models.

True AGI in science will require breakthroughs beyond current Transformer-based models, suggesting that long-term investment in foundational AI research into areas like memory and cumulative reasoning could yield disruptive advantages.

Source episodes

Timeline

Historical Context

Tao frequently references the history of science, particularly the work of Copernicus, Tycho Brahe, Kepler, and Newton, to frame the evolution of scientific discovery from data collection to theoretical explanation (Claims 9, 13, 14, 16).

2023

Makes the prediction that by 2026, AI will become a 'trustworthy co-author' in mathematics if used correctly (Claim 7).

Recent Months (pre-podcast)

Notes a surge in AI performance on specific mathematical challenges, with AI programs solving approximately 50 Erdős problems (Claim 12).

Current Period

Observes that progress on AI solving Erdős problems has paused, suggesting the 'low-hanging fruit' has been picked. He also states that he still uses pen and paper for the most difficult parts of math problems, as AI has not yet significantly accelerated this core task (Claims 1, 21).

Future (by 2026)

Reaffirms his 2023 prediction, believing AI is on track to be a trustworthy co-author in mathematics (Claim 7).

Future (within a decade)

Predicts that AI will be capable of automating many of the tasks that currently constitute the bulk of a mathematician's work (Claim 2).

Suggested prompts

How does Terence Tao reconcile his view that AI has a 1-2% success rate on math problems with the claim that an AI solved one of his own problems far faster than human experts? &nearr;What specific breakthroughs beyond current AI paradigms does Tao believe are necessary for AI to fully replace human mathematicians? &nearr;How does Tao's historical analysis of scientific revolutions, such as those involving Kepler and Newton, inform his predictions about the impact of AI on modern science? &nearr;What tools or processes does Tao envision for efficiently verifying the thousands of AI-generated theories, addressing the new bottleneck he identifies? &nearr;

Key concepts

AI in Mathematics 2 ep History of Science 1 ep AI Limitations 1 ep Human-AI Collaboration 1 ep Scientific Method 1 ep Erdős Problems 1 ep Kepler's Laws 1 ep Cryptography 1 ep Large Language Models 1 ep

Notable quotes

“So I think AI has basically driven the cost of idea generation down to almost zero.”

Terence Tao · Terence Tao – How the world’s top mathematician uses AI

“i i do believe that that hybrid um human plus ais will will dominate mathematics for a lot longer it it's it will require some additional breakthroughs beyond what we already have”

Terence Tao · Terence Tao – How the world’s top mathematician uses AI

“I think within a decade a lot of things that mathematicians currently do we spend a lot of the bulk of our time doing it and a lot of stuff we put in our papers today can be done by AI”

Terence Tao · Terence Tao – How the world’s top mathematician uses AI

“you made a prediction in 2023 that I think by 2026, what was it, it would be like a colleague in mathematics or? Yeah, a trustworthy co-author if used correctly. Which is looking pretty good in retrospect.”

Terence Tao · Terence Tao – How the world’s top mathematician uses AI

Report last updated: Apr 21, 2026

Get started free

Back to Entities Intelligence Report